You can run and orchestrate a robust MCP server that combines semantic reasoning, knowledge graphs, and neuro-symbolic querying. This server exposes compositional tools to decompose concepts, search proofs, query graphs, and manage knowledge with a scalable, language-agnostic architecture. It is designed to work with MCP clients over a standardized protocol, enabling you to build advanced AI workflows on top of RDF graphs and lambda-calculus representations.

How to use

Install and run the Congo River MCP Server locally, then connect your MCP client to the stdio-based runtime. You will decompose concepts, run graph queries, perform neuro-symbolic reasoning, and manage knowledge graphs through a single, compositional interface.

How to install

Prerequisites include Node.js 18+ and Python 3.10+. You also need a Supabase account (or a local PostgreSQL with pgvector) and API keys for Anthropic or OpenAI.

/home/mdz-axolotl/ClaudeCode/congo-river-mcp

Clone or navigate to directory

cd /home/mdz-axolotl/ClaudeCode/congo-river-mcp

Install Node dependencies

npm install

Install Python dependencies

pip install -r requirements.txt python -m spacy download en_core_web_sm

Configure environment

cp .env.example .env

Edit .env with your Supabase URL and API keys

Build TypeScript

npm run build

Initialize database

npm start -- --setup

Start server

npm start

## Additional sections

Configuration details and security notes help you tailor the server to your environment and keep sensitive information protected.

.env example

# Use Supabase
DB_TYPE=cloud
CLOUD_DB_URL=postgresql://postgres:[PASSWORD]@[PROJECT-REF].supabase.co:5432/postgres

# Add your API keys
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...

Add the MCP server configuration to your Claude Code project to enable the Congo River MCP Server as an active MCP endpoint.

mcp.json

{ "mcpServers": { "congo-river": { "command": "node", "args": ["dist/server.js"], "cwd": "/home/mdz-axolotl/ClaudeCode/congo-river-mcp", "type": "stdio", "env": { "TRANSPORT": "stdio", "DB_TYPE": "cloud", "CLOUD_DB_URL": "postgresql://...", "ANTHROPIC_API_KEY": "sk-ant-...", "OPENAI_API_KEY": "sk-..." } } } }

Start and monitor the MCP server through the standard MCP client workflow. The stdio configuration launches the Node-based server that serves MCP requests through the STDIO transport.

## Security and maintenance notes

Keep API keys out of version control by using environment variables and a secure secrets store. Regularly update dependencies and run health checks to ensure the knowledge graph, embeddings, and proofs remain consistent.

## Available tools

### triple\_decomposition

Decomposes concepts into RDF triples and stores them in the knowledge graph for later querying.

### lambda\_abstraction

Converts processes into lambda calculus, showing type signatures and applying beta reduction.

### proof\_search

Searches for proofs using forward/backward chaining, resolution, and returns proof trees.

### graph\_query

Queries the knowledge graph with SPARQL-like patterns and returns matching triples.

### neuro\_symbolic\_query

Hybrid reasoning that parses natural language to logical form, queries the graph, and returns grounded answers with proof traces.

### recommend\_language

Analyzes requirements and recommends optimal programming language with rationale.

### configure\_database

Manages database status, migrations, and configuration (local/cloud switching).

### export\_knowledge

Exports the knowledge graph to RDF or JSON for backup or portability.

### import\_knowledge

Imports triples into the knowledge graph from external sources.

### system\_status

Performs a comprehensive health check of the system and displays stats.